MindStat Spatial v1
R Engine
Skip to main content

Uses spatial coordinates from the Weights Builder. Build weights first.

Run Variogram first to fit model parameters, then click Run.

Run at least two outlier methods first, then click Generate.

Graph Neural Network (GNN)
GCN-based spatial regression with k-NN graph and Moran's I residual test
Data Source
Spatial Weights & Training
GNN Architecture
Train the model to see results
Or load sample data to get started
Spatial Autoregressive Neural Network (SAR-NN)
Joint optimisation of ρ (spatial lag coefficient) and a non-linear NN component for spatial prediction
Data & Spatial Weights
Architecture & Training
Train the model to see results
Spatial Convolutional Neural Network (Spatial CNN)
Convolutional kernels learn spatial patterns on a regular raster grid
Grid Configuration
CNN Architecture
Train the model to see results
Click to dismiss